Comparative Analysis and Bias Evaluation on Results of Some Items between Dry and Wet Biochemistry Analyzers

Objective To explore the clinical acceptability of the result deviation between two biochemical analysis systems (ROCHE P800 and Vitros 350 automatic biochemical analyzer) by methodology comparison, correlation analysis and bias evaluation. Methods According to the EP15-A issued by the National Committee on Clinical Laboratory Standards (NCCLS), ROCHE P800 (as the comparison analyzer) and Vitros 350 (as the laboratory analyzer) were used respectively to detect potassium (K﹢), natrium (Na﹢), chlorine (Cl﹣), calcium (Ca), glucose (Glu), urea nitrogen (Urea), creatine kinase (CK), creatinine (Cr) and aspartate aminotransferase (AST) of the serum. The results were analyzed and the bias were evaluated. The comparability of the two systems were judged according to the half of CLIA'88 (American Clinical Laboratory Improvement Amendment 88) permitted total error and verified relative bias value. Results Among these testing results, the correlation coefficients (r) of all items were higher than requirement specified by EP15-A (r ≥0.975). Except that the bias of urea was unacceptable (1/2 CLIA'88 permitted error and verified relative biase value), the biases of other items could be accepted in clinic between the two analytic systems. Conclusions If the same item is tested by different systems, it is necessary to judge the clinical acceptability by method comparison and bias estimation; and take ameliorative actions for incomparable results so as to ensure the comparability.